JPH0660899B2 - Water quality abnormality detector - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention provides image recognition of a plurality of aquatic animals bred using influent water and treated water of a water purification plant or a sewage treatment plant, The present invention relates to a water quality abnormality detection device that detects the presence or absence of a poisonous substance in test water from the behavior pattern of the above.

[Conventional technology]

At water purification plants, etc., in order to ensure the safety of the quality of the raw water that is taken in, some of the raw water is introduced to a water tank, where crucian carp, carp,
It breeds fish such as Japanese dace, eelfish, oikawa and goldfish. Similarly, fish may be bred to monitor the treated water, effluent water, river water, and lakes of sewage treatment plants for the presence of toxic substances in the water.

When a poisonous substance is mixed in water, these reared fish exhibit abnormal behavior or die, so that the poisonous substance is detected, and the abnormal behavior has been monitored visually before. However, since this cannot detect when a person is not monitoring, an automatic monitoring device has been developed. A conventional example is disclosed in Japanese Patent Laid-Open No. 46294/1986, and water quality is monitored by monitoring the behavior patterns of a plurality of organisms. That is, the temperature and the water quality are measured by the water quality sensor, and the normality / abnormality of the action pattern is determined with reference to the measured values. It is described that the action pattern at this time also includes speed and position.

Another prior application example is as follows. It is generally known that abnormal behavior of a fish, especially when it is against a poisonous substance, causes a nose-lifting phenomenon in which the mouth is exposed to the surface of the water, a rapid moving speed, and the like. The inventors have already devised a method for recognizing the behavior pattern of a school of fish by an image processing device in Japanese Patent Application No. 1985-45300 to detect these abnormal behaviors with high reliability. This is a method of recognizing the behavior pattern of a fish from the position and the moving speed in the depth direction in the aquarium. In addition, as a result of subsequent research, it was found that abnormal behavior of fish can be detected with a certain degree of accuracy in the depthwise position distribution.

Further, Japanese Patent Application No. 435/1987 describes a method of detecting an abnormality from the position of a fish that has been unable to withstand the water flow while the water flow is forcibly given to the fish.

[Problems to be Solved by the Invention]

The one disclosed in JP-A-61-46294 mentioned above is
The method of detecting behavior patterns is unclear. Japanese Patent Application Sho 62
In the one disclosed in No. 584 of the year, the nose-raising behavior of fish may occur by feeding food even if no poison is mixed,
Incomplete detection. The other disclosed in Japanese Patent Application No. 435/1987 is that the fish may be located on the downstream side due to a cause other than the contamination with poisonous substances, and the detection is also incomplete. In addition, this method divides the aquarium into three and breeds one fish in each, and detects the position of the center of gravity of the fish to detect the position and detect abnormalities. The method of recognizing is not specifically mentioned.

An object of the present invention is to provide a water quality abnormality detecting device capable of automatically detecting the positions of a plurality of fishes in one aquarium to detect water quality abnormality more reliably.

[Means for Solving the Problems]

The purpose of the above is to analyze the distribution of the fish school in the depth direction and the water flow direction by capturing the fish school that gives a water flow in the aquarium and raises it in a state where uniform water flows in the tank from the side with a camera installed horizontally. It is achieved by providing a means for

[Action]

The position distribution in the direction of water flow reduces the action of fish against water flow.
Since the position distribution in the depth direction indicates the presence or absence of the nose lifting behavior of the fish, if both are abnormal, it is possible to make a reliable determination without ambiguity if it is determined that the water quality is abnormal.

[Examples] Hereinafter, the present invention will be described in detail with reference to Examples. FIG. 1 is a block diagram showing an embodiment of the apparatus of the present invention. In the figure, the test water whose inflow is controlled by the regulating valve 2 attached to a part of the water supply pipe 1 is the supply pipes 3A, 3B, 3
It is led to the water tank 5 through C. Here, three water supply pipes were provided and a perforated threshold plate 4A was installed so that uniform water could flow into the water tank 5. Further, in this embodiment, the inflow control is performed by the adjusting valve, but the object can be achieved by using the method of changing the rotation speed of the water supply pump. The test water guided to the water tank 5 is drained from the drain pipe 8 via the threshold plate 4B. Inside the aquarium 5, there is a breeding space 9 partitioned by partition plates 4A, 4B, where fish 10A, 10B, 10C.
Breed. In the present embodiment, the case where there are three fish will be described, but the same embodiment can be applied to the case where there are more fish.

The illumination devices 6A and 6B are used for the fish 10 (fish 10A-1
0C is collectively called a fish, or a group of fish 10). A semitransparent plate 7 made of a semitransparent material such as frosted glass or paper is provided between the lighting devices 6A and 6B and the water tank 5. The translucent plate 7 scatters the light upon receiving the light from the illumination devices 6A and 6B, and the light emitted from the translucent plate 7 illuminates the aquarium 5.

An imaging device 20 such as an industrial television camera (ITV) is arranged on the opposite side of the water tank 5 as viewed from the lighting devices 6A and 6B.
That is, the light emitted from the lighting devices 6A and 6B and passing through the semitransparent plate 7 is input to the image pickup device 20 through the breeding space 9. This imaging sequence is shown in FIGS. 2 (a) and 2 (b).

The signal of the image pickup device 20 is guided to the image monitoring device 30. The details of the configuration and operation of the image monitoring device 30 will be described later, but the outline thereof is that captured images are captured at preset time intervals h, the fish body is image-recognized, and the position distributions in the water depth direction and the water flow direction are obtained. The behavior of the fish 10 is monitored and an alarm is issued in the case of an abnormality based on the monitoring result.

The monitor television 50 displays the captured image. The image monitor 60 receives the signal from the image monitoring device 30, and displays the image recognition result and the monitoring result such as the water depth direction position distribution and the water flow direction position distribution of the fish. The keyboard 70 inputs information for controlling the monitoring conditions of the image monitoring device 30 and the display of the CRT 80.

Next, the operation of the image monitoring device 30 will be described in detail. The timer 31S outputs a trigger signal for A / D conversion to the A / D converter 31 at every time interval h that is input by the keyboard 70 and initialized. This h is about 0.1 second to 2 seconds, and image processing is executed below at this time interval. Further, the timer 31S sets the image processing time h for one time and the number of times n of this image processing is repeated, and the measurement time T
(T = nh is obtained when image processing is repeated n times because one image processing time is h) so that a statistical behavior pattern of the school of fish during this period can be calculated. The measurement time T is about 10 seconds to 1 hour. The A / D converter 32 converts the multi-valued image signal from the imaging device 20 from an analog value to a digital value in synchronization with the trigger signal output from the timer 31S, and converts the digital multi-valued image signal into a multi-valued image memory 32M. Remember. The multi-valued image memory 32M has 256 vertical storage locations and 256 horizontal storage locations, and the luminance signal of the pixel corresponding to each storage location is stored as a digital value. I row and j column of this memory location (i = 1 to 256, j =
The signal (luminance) of 1st to 256th is represented as G (i, j). If the A / D converter 32 converts an analog value into a 7-bit digital value, G (i, j) has a digital value of 128 steps, and the multi-valued image memory 32M stores the multi-valued image An example is shown in FIGS. 3 (a) and 3 (b). FIG. 2 shows an image having multi-valued luminance. The luminance frequency distribution calculation circuit 33 calculates the luminance frequency distribution of the multivalued image. The luminance frequency distribution of FIG. 3 (a) is shown in FIG. 3 (b). The threshold value determination circuit 34 receives the calculation result of the luminance frequency distribution and determines the threshold value I. Next, a method of setting the threshold I will be described.

In the illumination method of the present invention, the school of fish 10 is always imaged as a dark object. Therefore, as shown in FIG. 3, only the area of the school of fish 10 (shown by hatching, and this area is f) is visited from the place where the brightness is low. A first threshold I ₁ is set there.
The area f is set by the fish body area setting circuit 34S. Since the area f differs depending on the state, it is set to the minimum area. This threshold setting method is especially effective when the water becomes cloudy. However, when the water is not turbid, it is better to use the threshold value shown in FIG. In FIG. 3B, the peak Pf represents the fish body, the peak Pb represents the background, and the portion Pe represents the gills and outline of the fish. Pf to extract only the fish
A second threshold value I ₂ is set at the boundary between P and Pe. As shown in FIG. 3, a threshold value is set to at least the brightness I ₁ in advance, and if there is a boundary (minimum value) between Pf and Pe while searching each frequency in the direction in which the brightness increases, I ₂ is set to this brightness. Choose.

Next, the binarization circuit 35 receives the signal of the multi-valued image memory 32M and the signal I (I ₁ or I ₂ ) of the threshold value determination circuit 34,
The multi-valued image is binarized and stored in the binary memory 35M. In this binarization, all the pixels whose brightness G (i, j) of the multi-valued image memory 32M is brighter than the threshold I are set to "0" level, and conversely, all the pixels darker than the threshold I are set to "1" level. By calculating all of these for each pixel, the background can be set to the “0” level and the school of fish 10 can be set to the “1” level. The result of binarizing FIG. 3 is shown in FIG.

The position distribution calculation circuit 36 receives the signal from the binary memory 35M and calculates the position distribution of the school of fish 10. The position distribution shown in FIG. 4 is defined by a distribution obtained by horizontally projecting the image shown in FIG. 4, as shown in FIG. That is, the position distribution of FIG. 5 represents at which water depth the fish school was in the water depth direction. That is,
The obtained position distribution is a distribution that represents the position of the school of fish in the depth direction. The water depth direction position distribution addition circuit 37 receives the result of the water depth direction position distribution calculation circuit 36 and accumulates the position distributions measured every time h so that an average position distribution can be calculated. This repetition is performed a predetermined number of times n based on the instruction from the timer 31S. The position distribution thus obtained normally has a peak at the bottom of the water tank as shown in FIG. 6 (a). The position distribution at the time of abnormality has a peak near the water surface as shown in FIG. 6 (b). This distribution represents so-called nose lift behavior.

In the water depth direction position distribution comparison circuit 38, the previously obtained position distribution at the normal time is input to the normal distribution setting circuit 38S and compared with the position distribution input from the water depth direction position distribution addition circuit 37. If it is determined to be abnormal, an alarm signal is sent to the alarm device 3
Input to 8A.

A method of comparing the position distributions in the water depth direction position distribution comparison circuit 38 will be described below. In order to evaluate the frequency of fish near the water surface, in FIG. 6, the ratio L of the area Ls near the water surface (shown by hatching) to the total area Lt of the distribution is shown.
Calculate s / Lt. In the depth direction position distribution comparison circuit 38, when Ls / Lt becomes larger than a predetermined value, it indicates that the fish 10 is taking a nose-raising action near the surface of the water, and thus it is determined to be an action abnormality. That is, it is determined that the water quality is abnormal. For example, if Ls / Lt is 0.2 or more, it is considered abnormal. The signal determined to be abnormal is input to the alarm device 38A.

On the other hand, the water flow direction position distribution calculation circuit 39 uses the binary memory 3
The position distribution in the water flow direction of the school of fish 10 is calculated by receiving the signal of 5M. The water flow direction position distribution of FIG. 7 is defined by the distribution obtained by photographing the image of FIG. 7 in the water flow direction, as shown in FIG. That is, the position distribution in the water flow direction in FIG. 8 represents where the fish school was located in the water flow direction. The water flow direction position distribution addition circuit 40 receives the result of the water flow direction position distribution calculation circuit 39 and accumulates the position distributions measured every time h to calculate an average position distribution. This repetition is repeated a predetermined number of times n based on the instruction of the timer 31S. The water flow direction position distribution thus obtained has a peak in the water flow upstream side as shown in FIG. 9 (a) under normal conditions. The position distribution at the time of abnormality is the 9th
The distribution has a peak on the downstream side as shown in FIG. This distribution represents the degree of weakness in which the school of fish 10 becomes intolerable to water flow. In the water flow direction position distribution comparison circuit 41, the position distribution obtained in advance in the normal state is the normal distribution setting circuit 41S.
To the position distribution input from the water flow direction position distribution adding circuit 40. If it is determined to be abnormal, it is input to the alarm device 42A.

A method of comparing the position distributions in the water flow direction position distribution comparison circuit 41 will be described below. The frequency of fish in the downstream area is evaluated in the same way as in the case of the distribution in the depth direction described above. In FIG. 9, when the ratio Ls / Lt of the area Ls of the downstream region to the entire distribution area Lt becomes larger than a predetermined value, it means that the fish 10 has weakened, and it is determined that the behavior is abnormal. For example, if Ls / Lt is 0.2 or more, it is considered abnormal. If it is determined to be abnormal, a signal is input to the alarm device 42A.

The alarm device 43A receives the signals from the alarm device 42A and the alarm device 38A, and when these signals are ON, determines that the behavior of the school of fish is abnormal.

〔The invention's effect〕

According to the present invention, abnormal behavior of a fish at the time of inflow of a poison is determined by the position in the water depth direction and the position in the water flow direction, so that there is no ambiguity as compared with the conventional method, and it is possible to detect with high reliability.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing an embodiment of the present invention, FIG. 2 is a diagram showing an image of a school of fish, FIG. 3 is a diagram showing a luminance frequency distribution, and FIGS. FIGS. 5 and 6 are views showing the position of the fish in the water depth direction, and FIGS. 8 and 9 are views showing the position of the fish in the water flow direction. 2 ... Regulator valve, 3 ... Water supply pipe, 5 ... Water tank, 6 ... Lighting device, 10 ... Fish, 20 ... Imaging device, 30 ... Image processing device, 32 ... A / D converter, 35 ... Binarization circuit,
36 ... Water depth direction position distribution calculation circuit, 39 ... Water flow direction position distribution calculation circuit, 38A, 42A, 43A ... Alarm device.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Kenji Baba 4026 Kuji Town, Hitachi City, Hitachi City, Ibaraki Prefecture Hitate Works, Ltd.Hitachi Laboratory Ltd. Inside Hitachi Research Laboratory (72) Inventor Mikio Yoda 52-1 Omika-cho, Hitachi City, Hitachi, Ibaraki Co., Ltd. Inside Omika Plant, Hitachi, Ltd. (72) Inventor Naoki Hara 5-2-1 Omika-cho, Hitachi City, Ibaraki Ceremony company Hitachi Ltd. Omika factory

Claims (1)

[Claims] 1. An aquarium for breeding aquatic animals, a water amount control means for adjusting the amount of water supplied in the aquarium, a diversion means for providing a uniform water flow in the aquarium, and a lateral view of the aquarium and aquatic animals. Image capturing means for capturing the image at a constant time interval and converting it into an electrical signal, and image information captured by the image capturing means with threshold values set so that the image of the aquatic animal can be distinguished from the image of the background portion. A binarizing means for binarizing, a water depth direction position distribution calculating means and a water flow direction position distribution calculating means for calculating the frequency distribution in the water depth direction and the water flow direction of the binary image showing the aquatic animals obtained by the means; Each of the cumulative amounts of the position distributions in the water depth direction and the water flow direction included in the shallow region and the downstream region that are predetermined using the calculation result of the water depth direction and water flow direction position distribution calculation means is the total of each direction. Area position Water abnormality detection apparatus characterized in that a determining means and the water quality abnormality when exceeds a predetermined rate for each cumulative amount of fabric.